tag:blogger.com,1999:blog-3524617892004055830.post4994725835204238845..comments2024-02-29T10:02:38.091+01:00Comments on Wiekvoet: Finding a jump in dataWingfeethttp://www.blogger.com/profile/01585623097384646816noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-3524617892004055830.post-89757449044453411932013-02-12T20:47:41.093+01:002013-02-12T20:47:41.093+01:00That sounds like the package I should have found. ...That sounds like the package I should have found. But I didn't. Thanks.Wingfeethttps://www.blogger.com/profile/01585623097384646816noreply@blogger.comtag:blogger.com,1999:blog-3524617892004055830.post-63167760293414458032013-02-12T10:16:19.772+01:002013-02-12T10:16:19.772+01:00As above there are several packages devoted to the...As above there are several packages devoted to the problem of detecting jumps in data. The one that springs to mind is strucchange for changes in regression. The advantage of this package is that you can enter many explanatory variables, just as you would to the lm function.Rebecca Killickhttps://www.blogger.com/profile/06801587920776862634noreply@blogger.comtag:blogger.com,1999:blog-3524617892004055830.post-41552453180585307602013-02-12T00:42:00.397+01:002013-02-12T00:42:00.397+01:00I like the use of tree regression a lot.
R also ...I like the use of tree regression a lot.<br /><br />R also has some very nice implementations of broken stick regression, which will give you a break in a linear trend.<br /><br />piecewise.linear()<br />bent.cable()<br /><br />Anonymoushttps://www.blogger.com/profile/07546021146056584277noreply@blogger.com